A properly implemented data layer
Your site explicitly pushes structured data to analytics at every key interaction not scraped from visible page text that breaks when someone redesigns a button.
Analytics
Marketing data getting harder to trust? You’re not alone
Marketers, it’s no surprise your data is confusing you. It’s not a you problem, it’s ten years of compounding complexity that’s causing the problem — and it’s about time someone wrote it all down in one place.
If you’ve ever sat in a reporting meeting watching three different dashboards show three different conversion numbers and quietly wondered whether any of them are right, you’re not alone, and you’re not doing it wrong.
Web data and analytics used to be pretty simple. You dropped a tag on your site, it counted visits, you got a number. Done. But between GDPR, cookie consent, platform migrations, browser privacy changes, and half a dozen “improvements” from Google over the years, the whole thing has become genuinely, legitimately complicated. Here’s how we got here.
Got a specific tracking question? We’ve answered the most common ones from marketers just like you.
A brief history of how it all got so messy
To understand where your setup should be today, it helps to understand the decade of disruption that got us here.
So, what does a solid setup actually look like now?
Think of your analytics setup like a plumbing system. GA4, your ad platforms, and your CRM are the taps you drink from but if the pipes behind the walls are leaking, corroded, or were installed by someone in a hurry in 2018, you’ll keep wondering why the water tastes funny. No amount of adjusting the tap fixes a pipe problem.
The good news is that the standards are now well established. Here’s what a properly built marketing data setup looks like today.
Must haves:
A regular analytics review, not just when things break
Most teams only look under the bonnet when a number goes suspiciously wrong. By then the damage is done, weeks or months of decisions made on bad data. Treating your analytics setup as something that needs periodic sense-checking (at minimum annually, and any time your site goes through significant change) is the difference between data you can trust and data you just hope is right.
Ecommerce event schema configured correctly
If you’re running an ecommerce site, your product data needs to match Google’s expected schema — item IDs, names, categories, prices all passed consistently across every event, from view_item through to purchase. Mismatches here silently corrupt your product performance reports, Shopping campaigns, and any audience built from purchase behaviour. It’s one of the most common things we find wrong in new client accounts.
Key conversion events verified
Every meaningful conversion in GA4 should be marked as a key event, tested, and confirmed to be firing not just assumed from the migration.
BigQuery export enabled
Raw data export means your historical data is safe if you ever need to reconfigure the property. Once gone, it can’t be recovered.
Consent Mode v2, configured correctly
Not just a cookie banner. Your consent signals must actually connect to GA4 and Google Ads, with modelled data recovery enabled for users who decline.
Should haves:
Enhanced conversions (Google + Meta)
Hashed first-party data sent back to ad platforms recovers attribution signal lost to browser restrictions. Meaningful uplift in reported conversions for most accounts.
Server-side tagging
Moves tracking off the browser and onto a server you control. More durable, more accurate, and less vulnerable to ad blockers and browser privacy updates.
The number one mistake marketers make with data
Trusting the number that agrees with the narrative.
When GA4, your ad platform, and your CRM all show different figures — and they always will — the temptation is to pick the one that supports the budget conversation you’re trying to have. But those discrepancies exist because each platform is measuring something genuinely different. GA4 measures browser behaviour. Ad platforms measure attribution within their own window. Your CRM measures what actually happened commercially. None of them are wrong; they’re just answering different questions.
The fix isn’t to pick a winner. It’s to build a reporting layer that uses each source for what it’s actually good at and stops letting inflated ad platform numbers drive budget decisions.
Not sure if your setup is actually working?
We run a (however many point from the software) point GA4 configuration audit that tells you exactly what’s missing, what’s misconfigured, and what it’s costing you in data quality. Most clients find at least two or three issues that have been silently affecting their reporting for months.
Want to know more?